Conventional electric energy can be easily adopted to a large scale by providing high quality electricity for wide-area transmissions. However, these energies are usually generated from exhaustible sources such as oil, natural gas, and coal, which are highly expensive in the long run and are the main causes of global warming. Meanwhile a large centralized energy system is more fragile and highly risky in countries like Japan where natural disasters occur frequently. A decentralized renewable energy system containing photovoltaic energy and wind power has been proposed as an alternative energy supply method. Within this system, the photovoltaic energy and wind power are well suited for the "local production and local consumption" with domestic energy transmission and are resilient to the unexpected disasters. The challenge of forming an optimal decentralized renewable energy system is to overcome its intrinsic disadvantages such as the instability and the limit of the power output. The research in this regard has drawn a lot of attention for the past twenty years. A decentralized renewable energy optimization problem is in principle categorized as nonlinear mixed integer programing problem(NMIP). Several challenging issues still remained in finding effective solution to NMIP through mathematical optimization. For instance, there is lack of reliable method to predict the energy generation and consumption; the weak scalability to large scale system is also existed due to the limited computing resource and the algorithm which are intrinsically not suitable for high speed computing. In this work, we report on employing the deep learning artificial intelligence techniques to predict the energy consumption and power generation together with the weather forecasting numerical simulation. The prediction and optimization are further examined by a small scale decentralized verification system (i-REMS) constructed inside the University campus. a novel optimization tool platform using Boltzmann machine algorithm for NMIP problem is also proposed for better computing scalable decentralized renewable energy system.